Detection of topological errors in distribution networks using state estimation residual patterns

Distribution system operators (DSOs) traditionally rely on incomplete and heterogeneous data (e.g., historical maps and GIS data) to reconstruct their network models, leading to inconsistencies in the network topology. With the widespread deployment of smart meters, DSOs now benefit from unprecedented data redundancy. With state estimation, DSO can exploit this redundancy and optimally combine measurements and network models to obtain the most likely system state. This study presents the idea of a simple but practical approach to detect probable errors, including topological issues and incorrect smart-meter phase assignments, based on the analysis of the geographical patterns of state-estimation residuals. The approach is demonstrated on an actual distribution network using real smart-meter measurements, and the practical insights gained from this implementation are discussed.